Fix the NaT fixture data

[jakirkham]

Fixes the NaT test case and fixture data by ensuring -2**63 is included at high frequency. The reason for this is explained by the code below.

In [1]: import numpy as np                                                      

In [2]: np.float64(-0.0).view("m8")                                             
Out[2]: numpy.timedelta64('NaT')

In [3]: np.float64(-0.0).view("m8").view(np.int64)                              
Out[3]: -9223372036854775808

In [4]: -2**63                                                                  
Out[4]: -9223372036854775808

TODO:

• Unit tests and/or doctests in docstrings
• tox -e py37 passes locally
• tox -e py27 passes locally
• Docstrings and API docs for any new/modified user-facing classes and functions
• Changes documented in docs/release.rst
• tox -e docs passes locally
• AppVeyor and Travis CI passes
• Test coverage to 100% (Coveralls passes)

[alimanfoo]

LGTM